As the relationship between land surface temperature (LST) and Normalized Difference Vegetation Index (NDVI) can not be able to
effectively apply after NDVI reached saturation[1]. In this study, the vegetation fraction, which derived from Dimidiate Pixel Model,
as the indicator of vegetation abundance has been introduced to analyze remote sensing of the change of urban heat islands (UHIs).
Landsat TM and ETM+ images of Wuhan from 1988 to 2002 were selected to retrieve the brightness temperatures and vegetation
fraction. Results show that, from 1988 to 2002, the intensity of UHI was increased and the average vegetation fraction of the whole
region was decreased. The urban ratio index (URI) of the study area was increased from 0.153 to 0.170 and the average vegetation
fraction decreased from 58.41% to 50.45%. Our analysis showed thatthe UHI effect has become moreprominent in areas of rapid
changes as its vegetation fraction decreasing in the past decade. The brightness temperature had strong correlation with vegetation
fraction,and the coefficient of determination value (R^2) was 0.867, 0.843, 0.841, 0.843 of 1988, 1991, 1996 and 2002.